Screening for bipolar disorder: the utility and comparative properties of the MSS and MDQ measures.

Background: Our objective was to further determine the diagnostic utility of the Mood Swings Survey (MSS) in distinguishing bipolar and unipolar disorders, and draw comparisons between this measure and the widely-used Mood Disorder Questionnaire (MDQ).
Methods: A total of 247 consecutively recruited patients attending the Black Dog Institute Depression Clinic were administered the Mood Swings Survey (MSS) as part of a computerized Mood Assessment Program (MAP), in addition to undergoing clinical assessment by two independent psychiatrists. The MDQ, along with a structured interview assessing DSM-IV criteria for bipolar disorder, was administered to a sub-sample of patients.
Results: The MSS-46 demonstrates comparable sensitivity and specificity to the MDQ (86.5% and 60.0% vs. 78.8% and 71.4%) when using pre-established cut-off scores. MSS diagnoses embedded within the computerized program correctly classified 82.2% of cases when compared to clinician diagnosis. Optimal cut-off scores derived in the current sample were > or = 35 (Se=88.5%, Sp=60.0%) for the MSS-46, and > or = 7 (Se=78.8%, Sp=71.4%) for the MDQ, indicating acceptable stability of cut-off scores in differing samples for both measures.
Limitations: ROC analyses compromised 'true' estimates of MSS sensitivity and specificity as a number of patients who did not affirm the initial screener question were excluded from these analyses.
Conclusions: Further work is required to evaluate the diagnostic utility of the MSS in differing clinical and community samples to determine the stability of its cut-off score and to refine the item set.